WCAG 2.2 checklist

WCAG groups criteria under four principles: content must be Perceivable, Operable, Understandable, and Robust (POUR). Each criterion has a conformance level: A (essential), AA (the standard legal target), or AAA (enhanced, rarely required site-wide). For compliance you target Level AA, which includes all of Level A.

• Text alternatives: every meaningful image, icon, and control has a text alternative (1.1.1).
• Colour & contrast: text meets 4.5:1 (3:1 for large text); do not rely on colour alone to convey meaning (1.4.1, 1.4.3, 1.4.11).
• Structure: text that looks like a heading is marked up as one, headings describe their section, and lists, tables and landmarks use correct markup (1.3.1, 2.4.6). Keeping heading levels in sequence is good practice, but a skipped level is not a WCAG failure on its own.
• Keyboard: everything works without a mouse, with no traps, and a visible focus indicator (2.1.1, 2.1.2, 2.4.7).
• Forms: every field has a label; errors are identified and described; required fields are marked (1.3.1, 3.3.1, 3.3.2).
• Navigation & links: link text makes sense out of context; pages have titles; focus order is logical (2.4.2, 2.4.4, 2.4.3).
• Media: captions for video, transcripts for audio (1.2.2, 1.2.1).
• New in 2.2: visible focus is not obscured (2.4.11), targets are at least 24x24px (2.5.8), no functionality needs dragging (2.5.7), and authentication does not rely on memory or puzzles (3.3.8).

Automated testing reliably verifies the machine-checkable criteria: contrast, alt-text presence, form labels, accessible names, page language, ARIA misuse and target size. That is roughly a third of the criteria by count, but those criteria account for a large share of the issues found on real pages, because the same few failures repeat everywhere. The rest, like whether alt text is meaningful or whether focus order makes sense, needs human judgement. A useful report keeps the two apart.

Use the free scan above to clear the automatable items first. You get every failing element outlined and mapped to its success criterion, plus the items flagged for manual review. Fix those, then do a keyboard and screen-reader pass for the human-judgement criteria.